Most teams searching for customer health score examples do not need another theory of weighting. They need concrete scorecards they can copy, adapt and defend: which metrics matter for an SMB account versus an enterprise account, how onboarding scoring differs from renewal scoring, and what "red" actually means in each case.
This guide gives four worked examples. For the step-by-step build behind any of them, see how to build a weighted customer health score. The examples below assume a 0–100 score with four bands: critical / rescue (0–24), at risk (25–49), medium / watch (50–74) and healthy (75–100).
How to Read These Examples
Each scorecard shows:
- Who it is for (segment or stage).
- Category weights that sum to 100%.
- The metrics inside each category, with direction and rough thresholds.
- What a red (critical or at-risk) score should mean operationally.
Treat the weights as starting points, not universal truth. Back-test them against your own renewals before you publish a model. Quivly's metrics guide covers how thresholds, direction and category combination work; use testing scores to calculate a draft against a real customer before you publish.
Example 1: Self-Serve / SMB Scorecard
Job of the score: catch silent churn and failed activation in a high-volume book where CSMs cannot live in every account.
Suggested weights
| Category | Weight | Why |
|---|---|---|
| Product usage | 45% | In self-serve, product behavior is the relationship |
| Support | 20% | Escalations and repeat tickets are early quit signals |
| Revenue / billing | 20% | Failed payments and downgrades often precede logo loss |
| Engagement | 15% | Sparse on purpose; most SMB accounts rarely talk to you |
Metric sketch
- Usage: weekly active users or key actions, 30-day trend, and whether the activation event ever fired. Higher is better for volume; a sharp negative trend should score worse than a stable low baseline that still hits the job-to-be-done.
- Support: open ticket count and resolution time. Lower is better. One unresolved severity ticket can outweigh quiet usage.
- Revenue: outstanding balance, recent downgrade, days until renewal. Lower outstanding balance is better.
- Engagement: days since last meaningful inbound (ticket, in-app message or email reply). Lower days-since is better, but do not over-weight silence in a product-led segment.
What red means here: the account is either not activated, trending hard down in usage, or stuck on a billing/support blocker. The play is usually automated recovery: activation nudge, payment retry, or a short human save for the highest ARR SMBs. Do not schedule a full QBR motion for every red SMB account.
Example 2: Mid-Market Scorecard
Job of the score: balance product adoption with relationship health for accounts that have a named CSM but limited executive access.
Suggested weights
| Category | Weight | Why |
|---|---|---|
| Product usage | 30% | Still the clearest proof of value |
| Engagement | 25% | CSM coverage exists; silence is actionable |
| Support | 20% | Mid-market churn often follows unresolved friction |
| Revenue | 15% | Renewal timing and contraction matter |
| Market signals | 10% | Hiring freezes, leadership changes, funding news |
Metric sketch
- Usage: breadth (active seats vs purchased) plus depth (core feature adoption) and 30–90 day trend.
- Engagement: calls in period, days since last call, Slack or email volume, and sentiment where available. Example threshold pattern from Quivly's metrics guide for days since last call with thresholds [14, 30, 60]: 0–14 healthy, 15–30 medium, 31–60 at risk, 60+ critical.
- Support: open tickets, first response time, resolution time.
- Revenue: MRR, outstanding invoices, days until renewal.
- Market signals: relevant company events in a 30- or 90-day lookback, interpreted in account context rather than as a fixed penalty.
What red means here: the CSM should open a save brief within a set SLA, confirm the primary risk reason with evidence, and either recover adoption or escalate to the account executive before the renewal window compresses. A red mid-market score without a named next step is just a badge.
Example 3: Enterprise Scorecard
Job of the score: protect multi-threaded, high-ACV accounts where politics, sponsors and delivery quality matter as much as raw usage.
Suggested weights
| Category | Weight | Why |
|---|---|---|
| Engagement | 30% | Sponsor access and multi-threading are survival skills |
| Product usage | 25% | Still required, but uneven usage across business units is common |
| Support / delivery | 20% | Incidents and stalled projects destroy executive trust |
| Revenue | 15% | Renewal timing, expansion posture, invoice health |
| Market signals | 10% | Reorgs, M&A, leadership changes, budget cycles |
Metric sketch
- Engagement: executive sponsor touch in the last 60 days, number of active stakeholders, QBR completion, days since last strategic call.
- Usage: adoption by business unit or workspace, not only a single company-wide total. A green global number can hide a red division that controls the renewal.
- Support / delivery: open severity tickets, aging blockers, implementation or expansion project slippage.
- Revenue: days until renewal, outstanding balance, contraction requests, competing RFP signals logged in CRM.
- Market signals: leadership change, layoff news, acquisition, or budget freeze affecting the buying center.
What red means here: executive-level risk. The play is a named rescue plan with sponsor mapping, risk reason, evidence and a dated exec touch. Enterprise red should rarely be handled only by an automated email.
Example 4: Onboarding Stage vs Renewal Stage
The same company can need two score lenses across the lifecycle. A single forever-model often punishes new accounts for not yet having mature usage, or ignores renewal politics because early activation looked fine.
Onboarding-stage emphasis (first 30–90 days)
| Category | Weight | Red means |
|---|---|---|
| Activation / time-to-value | 40% | Key milestone missed; no first value event |
| Project / delivery progress | 25% | Blocked tasks aging; customer-owned work stalled |
| Engagement | 20% | Kickoff attendees gone quiet; no admin identified |
| Support | 15% | Setup blockers unresolved |
Renewal-stage emphasis (last 90–120 days before renew)
| Category | Weight | Red means |
|---|---|---|
| Product usage trend | 30% | Sustained drop or narrow adoption |
| Engagement / sponsor health | 25% | No recent exec or champion contact |
| Support | 15% | Open severity or repeat themes |
| Revenue / commercial | 20% | Invoice issues, downgrade talk, competitor evaluation |
| Market signals | 10% | Buyer-side disruption |
Operationally, onboarding red should page the onboarding or implementation owner. Renewal red should page the CSM and AE with a save brief. If you only have one published model today, start there, then add a segment- or stage-specific configuration once you have enough history to justify it. Quivly supports health scoring configurations you can test on a single customer before publishing, which is the safest way to try a stage-specific model.
Shared Design Rules Across Every Example
- Prefer trends over raw totals when account size varies. Two hundred weekly actions can be excellent for SMB and weak for enterprise.
- Missing data is not zero. Exclude empty categories and rebalance, or the model will punish accounts with a delayed integration.
- Every red needs an owner and a play. A band without a response is decoration.
- Explain the score. Category contributions and evidence beat a mysterious number in a CRM field.
- Test before publish. Run the draft configuration on known-good, known-bad and recently churned accounts. Compare whether the new model would have flagged them earlier without flooding the team.
How Quivly Maps to These Examples
Quivly's Health Score is built around weighted categories such as engagement, product usage, support, revenue and market signals, with thresholds, direction and versioned configuration. The metrics guide documents available metrics by category and how category weights combine into the overall score. The testing flow lets you calculate a score for one customer on the published model or an unpublished draft, then inspect category contributions and metric details before you roll a change out.
Use the examples above to decide which model you need. Use the weighted-score build guide to implement it. Use testing on real accounts so the first version your CSMs see already matches how they triage.
Customer health score examples only help when they are specific enough to copy and strict enough to act on. Start with the segment that holds the most renewals at risk, publish one clear model, define what red means, and expand to stage-specific scoring once the first model earns trust.
